I really feel these kinds of studies are not conceptually well founded or informative in any actionable sense (especially if they ignore the fact the relevant gene expression may have been in development). - ThreadSky

wiringthebrain.bsky.social • 37 days ago

I really feel these kinds of studies are not conceptually well founded or informative in any actionable sense (especially if they ignore the fact the relevant gene expression may have been in development).

Reposted from Nature Neuroscience

Combined analysis of #GWAS and human brain single-cell transcriptomics data to infer the cell-types involved in the etiology of #schizophrenia and other psychiatric disorders 🧪🧠

www.nature.com/articles/s41...

Comments

stevepittelli.bsky.social•36 days ago

Especially since the GWAS results are mostly spurious correlations.

drlaramieduncan.bsky.social•32 days ago

We actually looked at developmental atlases too. The results suggest some very interesting distinctions between child and adult disorders… not published yet because we need better powered datasets for some of the analyses

wiringthebrain.bsky.social•37 days ago

Why should the pathogenesis be linked to some specific cell types? Why should the GWAS associations hit genes enriched in such hypothesised cell types? Why couldn't the pathophysiology be an emergent state of the whole system? (Which can be induced by all manner of insults)

wiringthebrain.bsky.social•37 days ago

What are we supposed to do with some marginal enrichments based on correlations of associations? Most genes are widely expressed in many cell types anyway. And showing enrichments to cortical and subcortical and excitatory and inhibitory neurons is not exactly narrowing things down...

drlaramieduncan.bsky.social•32 days ago

This is exactly what we say: cortical, subcortical, excitatory, and inhibitory are woefully broad terms. We found that prior studies didn’t describe all of the detail that they could have. We provide much greater detail here (though not in the stylized abstract, per journal preference)

wiringthebrain.bsky.social•37 days ago

The real take-home message is the LACK of convergence onto specific cell types, which fits with the lack of convergence onto specific biochemical processes or pathways

wiringthebrain.bsky.social•37 days ago

Which all reinforces the view that the pathology of psychiatric disorders like schizophrenia is better thought of as an emergent maladaptive attractor state of a dynamical system than a direct result of acute genetic dysfunction in some specific cells of the brain

wiringthebrain.bsky.social•37 days ago

(Unlike conditions like multiple sclerosis or Alzheimer's disease which DO have more proximal and specific cellular and molecular causes)

wiringthebrain.bsky.social•37 days ago

More thoughts on these issues here: If genomics is the answer, what's the question? http://www.wiringthebrain.com/2018/12/if-genomics-is-answer-whats-question.html

jknorrland.bsky.social•37 days ago

📌

richardsever.bsky.social•36 days ago

Is this a case where there's a pull to reductionism and gene-centric rather than true system level interpretation because of the experiments one can do (and the publishing rewards for doing so)?

wiringthebrain.bsky.social•36 days ago

Yes, IMO. We can do all the omics and then we can cross-correlate the results of all the different omics and generate "findings". And in some cases those are informative because the biology cooperates. But for psychiatric disorders, it doesn't...

iansudbery.bsky.social•36 days ago

I like the paradigm of things being maladaptive attractor states in dynamic systems. But I don't see why omics isn't/ couldn't be part of studying those states. It just requires analysis more subtle than a pvalue on an enrichment analysis.

tyrellturing.bsky.social•37 days ago

💯

It's amazing that the (very likely) possibility that psychiatric disorders are an emergent property of the whole network is not considered more often!

christineliu.art•36 days ago

!!!!!

markburgessosl.bsky.social•36 days ago

I believe it's part of a more general cultural flaw about thinking in terms of causes as "things" or objects rather than processes. Process thinking (spacetime thinking!) of where most answers lie, but that's not western ontology.

waully.bsky.social•31 days ago

In foundational ontologies, BFO, which is the basis for scientific ontologies, attributes causation to the things participating in the processes.
DOLCE on the other hand, can explicitly treat a process as the cause of another event or state. It is used more in business.
Funny how they diverge.

philipcball.bsky.social•37 days ago

This is a classic instance (I might even use it) of what I say in How Life Works about intervening at (and thus first identifying) the truly causal level(s). I felt I was probably stating the obvious, but apparently not.

wiringthebrain.bsky.social•37 days ago

No, clearly not obvious. There's an awful lot of technical (-omics) work being done in psychiatry just because it's possible, but without a well thought out conceptual basis.

philipcball.bsky.social•37 days ago

I think it's not just in psychiatry 😉

cedricboeckx.bsky.social•37 days ago

I agree that the developmental dimension is critical (hard to study, I know, but so incredibly important)

barakrotblat.bsky.social•36 days ago

my bet they will find increased entropy in psychiatric disease

scottwilliams9.bsky.social•36 days ago

Many of these studies are done because they can be. And the starry eyed are blown over by the mass of data. Question is always did this really add to our knowledge base or muddling with massive irrelevant dara

Comments

Posting Rules

Reply